My lab studies gene networks that control animal development and disease. Particular efforts focus on the control of segmentation and gastrulation in the early Drosophila embryo, the immune response in Drosophila larvae, and the differentiation of the notochord and heart in the sea squirt, Ciona intestinalis.
We are currently conducting a whole-genome analysis of gastrulation in the Drosophila embryo. This process is initiated by a maternal transcription factor called Dorsal, which is distributed in a broad concentration gradient across the dorsal-ventral axis of the early embryo. High levels of the gradient initiate the differentiation of the mesoderm, while low levels control the development of the neurogenic ectoderm and dorsal ectoderm. An extensive microarray screen identified as many as 40 new Dorsal target genes, including a Rho GTPase, an FGF-like signaling molecule, an ADAM metalloprotease, and a tumor necrosis factor (TNF). Some control the invagination of the mesoderm through the ventral furrow, while others are essential for the spreading of the mesoderm along the internal surface of the neurogenic ectoderm. It is possible to describe gastrulation in terms of threshold readouts of the Dorsal gradient. High levels of Dorsal activate target genes required for the invagination and spreading of the mesoderm, such as the FGF receptor Heartless. In contrast, low levels of the gradient activate secreted signaling molecules in the neurogenic ectoderm that control the behavior of the neighboring germ layers.
There are ~50 Dorsal target genes in the Drosophila genome, and a variety of bioinformatics methods allowed us to isolate enhancers for 15 of these genes. Coordinately regulated enhancers that respond to similar thresholds of the Dorsal gradient contain shared sequence motifs. Dorsal target enhancers are typically 500 bp in length and contain clustered binding sites for transcriptional activators and repressors (including binding sites for Dorsal itself). Simple clustering of binding sites is not sufficient to unravel a "cis-regulatory code", which links primary DNA sequence to differential gene activity. The elucidation of a cis-code depends on "grammar", that is, a fixed organization of cis-regulatory elements, such as helical phasing between two adjacent activator sites that work in a synergistic fashion to stimulate transcription. We are obtaining evidence that a fixed grammar governs the immune response in Drosophila larvae. Many immunity genes contain closely linked binding sites for Rel and GATA transcription factors. The binding sites are organized in the same relative orientation, and inverting the sites causes a severe reduction in the immune response.
The sea squirt, Ciona intestinalis, is a simple chordate with a small genome that has been recently sequenced and assembled. The Ciona genome represents a simplified version of vertebrate genomes. Large gene families in vertebrates are represented by just a few members in Ciona. For example, there are 22 FGF genes in a typical vertebrate, but only 6 in Ciona, and each of these corresponds to a subfamily of 3 or 4 vertebrate genes. The Ciona tadpole is initially composed of just ~1,000 cells and there is complete lineage information. It is possible to introduce transgenic DNAs into developing embryos using simple electroporation methods. For this reason we have been able to isolate large numbers of tissue-specific enhancers, including those that mediate gene expression in the notochord and neural tube. Bioinformatics methods are being used to identify shared sequence motifs within coordinately regulated enhancers. We hope to use this information to identify and characterize orthologous enhancers in vertebrates.